EP0822503A1 - Système de recouvrement de documents - Google Patents

Système de recouvrement de documents Download PDF

Info

Publication number
EP0822503A1
EP0822503A1 EP97113355A EP97113355A EP0822503A1 EP 0822503 A1 EP0822503 A1 EP 0822503A1 EP 97113355 A EP97113355 A EP 97113355A EP 97113355 A EP97113355 A EP 97113355A EP 0822503 A1 EP0822503 A1 EP 0822503A1
Authority
EP
European Patent Office
Prior art keywords
feature
document
input
classification
storage unit
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP97113355A
Other languages
German (de)
English (en)
Inventor
Masako Nomoto
Naohiko Noguchi
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Panasonic Holdings Corp
Original Assignee
Matsushita Electric Industrial Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Matsushita Electric Industrial Co Ltd filed Critical Matsushita Electric Industrial Co Ltd
Publication of EP0822503A1 publication Critical patent/EP0822503A1/fr
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data

Definitions

  • the present invention relates to a document retrieval system for retrieving a document suitable for a user's intention of retrieval from electronic documents.
  • a retrieval system based on exact match technique in which a user inputs queries comprising character strings and logic operators and obtains a document assembly, which satisfy the condition
  • a retrieval system based on partial match technique in which the similarity between the input by the user and the document to be retrieved are compared and collated with each other by some measure, and the documents are ranked according to the closeness to the user's intention of retrieval.
  • the retrieval based on the exact match technique is advantageous in that the query inputted by the user clearly corresponds to the document assembly as the result of retrieval.
  • the vocabulary used in the document to be retrieved is often unclear to the user, and it is difficult to specify a suitable keyword in the query, and in case a large amount of document assemblies have been obtained as the result of retrieval it is so hard for a user to choose relative documents one by one.
  • the retrieval based on the partial match technique is advantageous in that ranking is assigned to the documents as the result of retrieval according to the closeness to the user's intention of retrieval, while it is not necessarily clear to the user as to what kind of document are ranked on what basis.
  • linguistic features expressing content of query and document are classified into concept, which expresses each content, to clarify the query and the document, and the concepts which comprise linguistic feature is shown of the document to be retrieved actually corresponding to each concept, which comprises the query and document, and to support in efficiently selecting the document, which is closer to the user's intention of the retrieval.
  • the document retrieval system comprises an input/output control unit for receiving input from a user and for showing result of processing to the user, a user request processing unit for receiving a request other than query among the inputs from the user and for processing content of the request, a document storage unit for storing a document to be retrieved, a feature extracting unit for extracting linguistic feature of the query inputted from said input/output control unit as input feature or for extracting linguistic feature of the document stored in the document storage unit as document feature, an input feature storage unit for storing input feature extracted from the by the feature extracting unit, a document feature storage unit for storing linguistic feature taken out from the document by the feature extracting unit, a feature classifying unit for classifying the input feature stored in the input feature storage unit to correspond to concept of the content expressed by the query or for classifying the document feature stored in the document feature storage unit to correspond to concept of the content expressed by the document, an input feature classification storage unit for storing correspondence between the input feature classification and the input feature as the
  • the user can confirm easily the user which concept the query or the document is comprised of, and further which linguistic feature corresponds to each concept, and hence, efficiently select the document, which is closer to the intention of retrieval.
  • the system according to Claim 1 of the present invention comprises an input/output control unit for receiving input from a user and for presenting result of processing to the user, a user request processing unit for receiving a request other than query among the inputs from the user and for processing content of the request, a document storage unit for storing a document to be retrieved, a feature extracting unit for extracting linguistic feature of the query inputted from said input/output control unit as input feature or for extracting linguistic feature of the document stored in the document storage unit as document feature, an input feature storage unit for storing input feature extracted from the query by the feature extracting unit, a document feature storage unit for storing linguistic feature taken out from the document by the feature extracting unit, a feature classifying unit for classifying the input feature stored in the input feature storage unit to correspond to partial concept of the content expressed by the input statement or for classifying the document feature stored in the document feature storage unit to correspond to concept of the content expressed by the document, an input feature classification storage unit for storing correspondence between the input feature classification and the input feature as the
  • the input feature classification stored in the input feature classification storage unit and the correspondence between the input feature classification and the input feature of the specified document and the document feature classification stored in the document feature classification storage unit and the correspondence between the document feature classification and the document feature, and further, the document feature stored in the document feature storage unit are added, deleted or corrected, whereby the contents of the input statement and the document are expressed more adequately.
  • the system according to Claim 3 of the present invention comprises a feature collating unit for collating an input feature classification stored in the input feature classification storage unit and the corresponding input feature with the document feature classification stored in the document feature classification storage unit and the corresponding document feature, and a collating result storage unit for storing collating method and result by the feature collating unit, whereby the contents of the query and the document are compared and collated with the concept expressed by the input feature classification and the input feature belonging to the concept and the concept expressed by the document feature classification and the document feature belonging to the concept, and, as the result of calculation of similarity between the query and the document, not only the score and the ranking of each of the documents finally obtained but also method and result of collating of the similarity are presented to the user, thereby demonstrating to the user at which viewpoint the collating has been performed and how each document has been evaluated.
  • the feature collating unit collates them with the document feature classification and the document feature of each document, and the result of collating is presented to the user, whereby, by regarding the specific document feature classification and the document feature provisionally as the input feature classification and the input feature and collecting them with the document, if there is any concept or linguistic feature not appearing in the original input feature classification or the input feature and being regarded as adequate as the input feature classification or the input feature, these are collated with the document and the effect can be easily confirmed by the user.
  • the feature collating unit collates the input feature classification or the input feature thus weighted with the document feature classification or the document feature of each document, and the result of collating is presented to the user, whereby the user gives the degree of importance to the input feature classification or the input feature as a concept to constitute the input statement or as linguistic feature, to clearly demonstrate the intention of retrieval, and accuracy of retrieval is increased.
  • Fig. 1 is a block diagram showing functional arrangement of a document retrieval system according to an embodiment of the present invention.
  • reference numeral 11 represents an input/output control unit
  • 12 represents a user request processing unit
  • 13 is a data storage unit
  • 14 is a document feature storage unit
  • 15 is an input feature storage unit
  • 16 is a document feature classification storage unit
  • 17 is an input feature classification storage unit
  • 18 is a collating result storage unit
  • 19 is a document storage unit
  • 20 is a feature classifying unit
  • 22 is a feature collating unit.
  • a query described in natural language is inputted from a user via the input/output control unit 11.
  • the feature extracting unit 20 analyzes the input statement, extracts important words and phrases as linguistic features, and if necessary, these are stored in the input feature storage unit 15 together with statistical information such as frequency of appearance of these features and degree of importance . It is also possible to describe these input features by developing them to homonyms, synonyms, narrower term, etc. using a thesaurus.
  • Fig. 2 shows examples of a query.
  • Fig. 3 summarizes examples of data stored in the input feature storage unit 15 in case unnecessary words such as symbols, particles, etc. are exempted from the words obtained through morphological analysis to the query of Fig. 2 and the remaining words are regarded as input features.
  • the feature extracting unit 20 analyzes each of the documents stored in the document storage unit 19 in similar manner. Important words and phrases are extracted as linguistic features, and if necessary, these are stored in the document feature storage unit 14 together with statistical information such as frequency of appearance of these features and degree of importance.
  • the feature extracting unit 20 extracts important words and phrases, for example, information such as frequency of the words, distribution of words among documents, parts of speech, appearing position in the document, syntactic and semantic relationship with other words, etc. are used for the judgment of the degree of importance of the words.
  • the feature classifying unit 21 classifies the input features stored in the input feature storage unit 15 to correspond to each concept, which comprises the query, puts a classification name and stores it to the input feature storage unit 17.
  • the feature classifying unit 21 possesses hierarchical thesauruses. Nodes at a given depth on the hierarchical thesauruses are set as criteria for classification, and the words having semantically closer concept under the node are put together. Also, there is a method to put together specific words having closer syntactic or semantic relationship using concurrence dictionary or concept dictionary.
  • Fig. 4 gives an example of a part of the hierarchical thesaurus possessed by the feature classifying unit 21.
  • a word with a concept corresponding to the node at a given depth in hierarchical thesaurus possessed by the feature classifying unit 21 with respect to the input feature of a certain group and with broader concept of a plurality of words is used.
  • Fig. 5 shows examples of data stored in the input feature classification storage unit 17 such as input feature corresponding to the input feature classification obtained using the broader concept 1 as criterion for classification in the hierarchical thesaurus of Fig. 4.
  • the feature classifying unit 21 also classifies the document feature stored in the document feature storage unit 14 so that it corresponds to concept comprising the document, puts a classification name and stores it in the document feature classification storage unit 16.
  • the method to classify the document feature and to determine the classification name is the same as in the classification of the above input feature.
  • Fig. 6 shows examples of data stored in the document feature classification storage unit 16 such as correspondence between the document feature classification and the document feature.
  • the document feature classification common to the input feature classification and the document feature common to the input feature are enclosed by the symbol "[]", and identifiers such as "A", "B", etc. are put to the input feature classification and the document feature classification.
  • the feature collating unit 22 collates the document feature classification and the corresponding document feature stored in the document feature classification storage unit 16 with the input feature classification and the corresponding input feature stored in the input feature classification storage unit 17 and determines the ranking of the documents.
  • the results of the collating of the document and a document ranking table indicating the ranking are stored in the collating result storage unit 18, and the document ranking table is presented to the user via the input/output control unit 11.
  • Score E ( ⁇ ) of the document ⁇ ⁇ (Weight of document feature classification to which document feature belongs x Weight of document feature x Frequency of appearance of the document feature in the document ⁇ )
  • the document feature classification and the document feature are evaluated depending upon whether the corresponding input feature classification and input feature are present or not.
  • weight is given according to:
  • Fig. 7 shows examples of the document ranking table in the above case.
  • subtotal of the scores of document feature of each document is given as partial score, and the same applies to all of the subsequent ranking tables.
  • frequency of appearance of the document feature in each document or weight given to the document feature classification and the document feature can be given in the ranking table.
  • ranking can be given by putting weight to the input feature classification and the input feature.
  • the user specifies, via the user request processing unit 12, thus collating with the documents having the document feature classification and the document feature shown in Fig. 6, presuming that the weight of the input feature classification is 1, the weight of the input feature belonging to the input feature classification C is 1, and the weight of the other input feature classification and the input feature is 0.
  • Fig. 8 shows examples of the data of the corrected input feature classification storage unit 17 as the result of the weighting to the data of the input feature classification storage unit of Fig. 5 at the request of the user.
  • Fig. 9 summarizes an example of document ranking table obtained as the result.
  • the document feature classification and the document feature with weighting are enclosed by the symbol "[ ]”.
  • Fig. 10 shows examples of the provisional input feature classification and the input features to be stored in the input feature classification storage unit 17 in case the weight of the provisional input feature classification C is 1 and weight of each of the provisional input features "boil", “steam” and “dress” is 1 respectively.
  • the weights of the document feature classification and the document feature are determined according to the above Rule 2.
  • the ranking of the documents is calculated using the evaluation function of the equation 1, supposing that the indocument frequency of the document feature of each document is 1, the ranking is: document 3 - document 2 - document 1.
  • the ranking table thus obtained is given in Fig. 11.
  • the user can add the document feature classification having no corresponding input feature classification or the document feature having no corresponding input feature, as an input feature classification or an input feature.
  • the user inputs a request via the user request processing unit 12, that the document features "boil", “steam” and "dress” belonging to the document feature classification C of the documents 2 and 3 in Fig. 6 should be added to the input feature classification C of Fig. 5. Because the above three input features specified by the user are not included in the input feature classification C, among the document features of the document feature classification C, these are added to the input feature classification storage unit 17 as new input features of the input feature classification C.
  • Fig. 12 shows an example of data of the input feature classification storage unit 17 after correction, which is obtained by adding the document feature classification F and its document feature and some of the document features among the document feature classification C to the data of the input feature classification storage unit of Fig. 5.
  • the newly added input feature classification and the input features newly added are enclosed by the symbols "* *".
  • ranking is given by newly adding weight to the corrected input feature classification or the input features.
  • the user sees the data of the input feature classification storage unit 17 corrected as shown in Fig. 12 and wants to retrieve by selecting the document relating to "stewing vegetables or fishes".
  • the user sets the weight of the input feature classifications A and F of Fig. 12 as 5, the weight of input feature belonging to the input feature classifications A or F as 1, the weight of the input feature classification C as 5, the weight of the feature "stew” among the input features belonging to the input feature classification C as 10, the weight of the input features other than "stew” belonging to the input feature classification C as 0, and the weight of the other input feature classifications and the input features as 0.
  • Fig. 13 shows the data of the input feature classification storage unit 17 thus corrected.
  • the weight of the document feature classification and the document feature are determined according to the above Rule 2. If it is supposed that the in-document frequency of the document feature of each document is 1, the ranking of the documents is calculated using the evaluation function of the equation 1, and the ranking is: document 2 - document 3 - document 1. Fig. 14 shows the document ranking table thus obtained.
  • the contents of query and the document are presented to the user as corresponding relationship between the partial concept expressing each of the contents and linguistic features belonging to each concept. If necessary, the user performs collating by adding correction and weighting to each of the concepts and linguistic features, and the collating method and the result of collating are easily confirmed. As a result, it is possible to obtain advantageous effects, i.e. to efficiently select the document closer to the user's intention of retrieval and to perform retrieval with high accuracy.
  • Linguistic features extracted from the query are classified into concepts expressing content of the query, and linguistic features extracted from the document are classified into concepts expressing content of the document.
  • the user confirms of which concept the input statement and the document are comprised and which kind of linguistic feature corresponds to each concept, and the system assists the user to adequately select a document, which is closer to the intention of retrieval.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
EP97113355A 1996-08-02 1997-08-01 Système de recouvrement de documents Withdrawn EP0822503A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP20455796A JP3198932B2 (ja) 1996-08-02 1996-08-02 文書検索装置
JP204557/96 1996-08-02

Publications (1)

Publication Number Publication Date
EP0822503A1 true EP0822503A1 (fr) 1998-02-04

Family

ID=16492461

Family Applications (1)

Application Number Title Priority Date Filing Date
EP97113355A Withdrawn EP0822503A1 (fr) 1996-08-02 1997-08-01 Système de recouvrement de documents

Country Status (2)

Country Link
EP (1) EP0822503A1 (fr)
JP (1) JP3198932B2 (fr)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000005663A2 (fr) * 1998-07-24 2000-02-03 Jarg Corporation Systeme de base de donnees d'ordinateur repartie et procede de mise en oeuvre d'une recherche d'objets
WO2000005664A1 (fr) * 1998-07-24 2000-02-03 Jarg Corporation Systeme et procede de recherche fondes sur de multiples ontologies
WO2000055765A1 (fr) * 1999-03-05 2000-09-21 Cai Co., Ltd. Procede de tri/recherche/resume de documents
WO2001022279A1 (fr) * 1999-09-24 2001-03-29 France Telecom Procede de classification thematique de documents, module de classification thematique et moteur de recherche incorporant un tel module
GB2362004A (en) * 2000-04-19 2001-11-07 Glenn Courtney Smith Data object matching using a classification index
US6598043B1 (en) 1999-10-04 2003-07-22 Jarg Corporation Classification of information sources using graph structures
WO2004010324A2 (fr) * 2002-07-19 2004-01-29 Go Albert France Sarl Systeme d'extraction d'informations dans un texte en langage naturel
WO2004114162A2 (fr) * 2003-06-17 2004-12-29 Google, Inc. Categorisation de demandes pour recherche de listes d'adresses commerciales

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11259504A (ja) * 1998-03-11 1999-09-24 Mitsubishi Electric Corp データベース作成装置およびデータベース検索装置
US6757692B1 (en) * 2000-06-09 2004-06-29 Northrop Grumman Corporation Systems and methods for structured vocabulary search and classification
KR102149844B1 (ko) * 2015-09-09 2020-08-31 엘에스엠트론 주식회사 리셉터클 커넥터용 미드플레이트 및 리셉터클 커넥터
KR101697264B1 (ko) 2015-12-18 2017-01-17 한국단자공업 주식회사 Usb 플러그 커넥터의 접지 구조, usb 플러그 커넥터, usb 플러그 케이블장치, 및 usb 커넥터 어셈블리
JP6538134B2 (ja) * 2017-10-05 2019-07-03 日本航空電子工業株式会社 コネクタ
KR102149848B1 (ko) 2020-07-30 2020-08-31 엘에스엠트론 주식회사 리셉터클 커넥터용 미드플레이트 및 리셉터클 커넥터

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0638870A1 (fr) * 1985-03-27 1995-02-15 Hitachi, Ltd. Procédé de recouvrement d'informations
WO1995030981A1 (fr) * 1994-05-05 1995-11-16 Hutson William H Procede et systeme d'analyse d'informations textuelles en temps reel
EP0704810A1 (fr) * 1994-09-30 1996-04-03 Hitachi, Ltd. Procédé et dispositif pour classer des informations de documents

Family Cites Families (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2839555B2 (ja) * 1989-06-14 1998-12-16 株式会社日立製作所 情報検索方法
JPS63157228A (ja) * 1986-12-22 1988-06-30 Nec Corp 階層形ブラウジング方式
JPH07120357B2 (ja) * 1987-07-27 1995-12-20 松下電器産業株式会社 文書検索装置
JPS6446831A (en) * 1987-08-17 1989-02-21 Nippon Telegraph & Telephone Automatic key word extracting device
JPH0250262A (ja) * 1988-08-12 1990-02-20 Canon Inc 情報登録検索装置
JPH02178877A (ja) * 1988-12-29 1990-07-11 Nec Corp 対話制御方式
JPH02235176A (ja) * 1989-03-09 1990-09-18 Ricoh Co Ltd 概念検索装置
JP2840664B2 (ja) * 1989-04-27 1998-12-24 株式会社リコー 意味分類方法
JP2885487B2 (ja) * 1990-07-26 1999-04-26 日本電信電話株式会社 文書内情報検索装置
JP2943447B2 (ja) * 1991-01-30 1999-08-30 三菱電機株式会社 テキスト情報抽出装置とテキスト類似照合装置とテキスト検索システムとテキスト情報抽出方法とテキスト類似照合方法、及び、質問解析装置
JPH0644326A (ja) * 1991-10-21 1994-02-18 Nec Home Electron Ltd 情報検索装置
JPH05151273A (ja) * 1991-11-29 1993-06-18 Nec Corp シソーラス検索システム
JPH05250411A (ja) * 1992-03-09 1993-09-28 Nippon Telegr & Teleph Corp <Ntt> 検索条件式作成装置
JPH05334364A (ja) * 1992-06-03 1993-12-17 Nippon Telegr & Teleph Corp <Ntt> 検索条件式作成方法
JPH06274541A (ja) * 1993-03-17 1994-09-30 Nippon Steel Corp 文献検索システム
JPH0793345A (ja) * 1993-09-20 1995-04-07 Toshiba Corp 文書検索装置
JPH086963A (ja) * 1994-06-15 1996-01-12 Fuji Xerox Co Ltd 文字列検索装置
JPH08171569A (ja) * 1994-10-28 1996-07-02 Fuji Xerox Co Ltd 文書検索装置
JPH08137841A (ja) * 1994-11-04 1996-05-31 Canon Inc 文書処理装置及び方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0638870A1 (fr) * 1985-03-27 1995-02-15 Hitachi, Ltd. Procédé de recouvrement d'informations
WO1995030981A1 (fr) * 1994-05-05 1995-11-16 Hutson William H Procede et systeme d'analyse d'informations textuelles en temps reel
EP0704810A1 (fr) * 1994-09-30 1996-04-03 Hitachi, Ltd. Procédé et dispositif pour classer des informations de documents

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
GIGER H P: "CONCEPT BASED RETRIEVAL IN CLASSICAL IR SYSTEMS", PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL. (SIGIR), GRENOBLE, JUNE 13 - 15, 1988, no. CONF. 11, 13 June 1988 (1988-06-13), CHIARAMELLA Y, pages 275 - 289, XP000295044 *

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000005664A1 (fr) * 1998-07-24 2000-02-03 Jarg Corporation Systeme et procede de recherche fondes sur de multiples ontologies
WO2000005663A3 (fr) * 1998-07-24 2000-04-27 Jarg Corp Systeme de base de donnees d'ordinateur repartie et procede de mise en oeuvre d'une recherche d'objets
WO2000005663A2 (fr) * 1998-07-24 2000-02-03 Jarg Corporation Systeme de base de donnees d'ordinateur repartie et procede de mise en oeuvre d'une recherche d'objets
WO2000055765A1 (fr) * 1999-03-05 2000-09-21 Cai Co., Ltd. Procede de tri/recherche/resume de documents
WO2001022279A1 (fr) * 1999-09-24 2001-03-29 France Telecom Procede de classification thematique de documents, module de classification thematique et moteur de recherche incorporant un tel module
FR2799023A1 (fr) * 1999-09-24 2001-03-30 France Telecom Procede de classification thematique de documents, module de classification thematique et moteur de recherche incorporant un tel module
US7003519B1 (en) 1999-09-24 2006-02-21 France Telecom Method of thematic classification of documents, themetic classification module, and search engine incorporating such a module
US6772148B2 (en) 1999-10-04 2004-08-03 Jarg Corporation Classification of information sources using graphic structures
US6598043B1 (en) 1999-10-04 2003-07-22 Jarg Corporation Classification of information sources using graph structures
GB2362004A (en) * 2000-04-19 2001-11-07 Glenn Courtney Smith Data object matching using a classification index
WO2004010324A2 (fr) * 2002-07-19 2004-01-29 Go Albert France Sarl Systeme d'extraction d'informations dans un texte en langage naturel
WO2004010324A3 (fr) * 2002-07-19 2004-04-01 Albert Inc S A Systeme d'extraction d'informations dans un texte en langage naturel
EP1391830A1 (fr) * 2002-07-19 2004-02-25 Albert Inc. S.A. Système d'extraction d'informations dans un texte en langage naturel
US8170867B2 (en) 2002-07-19 2012-05-01 Go-Albert France System for extracting information from a natural language text
WO2004114162A2 (fr) * 2003-06-17 2004-12-29 Google, Inc. Categorisation de demandes pour recherche de listes d'adresses commerciales
WO2004114162A3 (fr) * 2003-06-17 2005-03-03 Google Inc Categorisation de demandes pour recherche de listes d'adresses commerciales

Also Published As

Publication number Publication date
JP3198932B2 (ja) 2001-08-13
JPH1049543A (ja) 1998-02-20

Similar Documents

Publication Publication Date Title
Witten Text Mining.
US7870118B2 (en) Search system
US5937422A (en) Automatically generating a topic description for text and searching and sorting text by topic using the same
US20040049499A1 (en) Document retrieval system and question answering system
US20050102251A1 (en) Method of document searching
EP1661031A1 (fr) Systeme et methode pour traiter un texte au moyen d&#39;une suite de techniques de desambiguisation
Janik et al. Wikipedia in action: Ontological knowledge in text categorization
EP1606704A2 (fr) Systemes et procedes visant a affiner une demande de recherche interactive
EP0822503A1 (fr) Système de recouvrement de documents
Ramprasath et al. A survey on question answering system
King et al. Mining world knowledge for analysis of search engine content
Li et al. The role of semantic information in learning question classifiers
Gopan et al. Comparative study on different approaches in keyword extraction
Krishnan et al. A supervised approach for extractive text summarization using minimal robust features
KR20230077589A (ko) 외부 데이터베이스를 활용하여 금융 도메인의 다양한 질의에 대한 의도를 분류하고 및 답변을 검색하는 방법 및 시스템
Mima et al. The ATRACT workbench: Automatic term recognition and clustering for terms
Strzalkowski Natural language processing in large-scale text retrieval tasks
KR100407081B1 (ko) 문서 검색 및 분류 방법 및 장치
Basili et al. A robust model for intelligent text classification
Manjula et al. Semantic search engine
Fujii et al. Toward the automatic compilation of multimedia encyclopedias: associating images with term descriptions on the web
Lin et al. Chinese Question Classification Using Alternating and Iterative One-against-One Algorithm.
Golub Using controlled vocabularies in automated subject classification of textual web pages, in the context of browsing
Szczepaniak et al. Practical evaluation of textual fuzzy similarity as a tool for information retrieval
Basili et al. Empirical investigation of fast text classification over linguistic features

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 19970801

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): DE FR GB

AX Request for extension of the european patent

Free format text: AL;LT;LV;RO;SI

AKX Designation fees paid

Free format text: DE FR GB

RBV Designated contracting states (corrected)

Designated state(s): DE FR GB

17Q First examination report despatched

Effective date: 20010913

17Q First examination report despatched

Effective date: 20010913

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 20070301